组合预测在医药流通企业销售预测中的应用研究

发布时间：2018-06-11 23:33

本文选题：销售预测 + 组合预测　；参考：《华南理工大学》2015年硕士论文

【摘要】：药品的销售规律并不是一个简单的线性行为方式,它更是受市场、经济、政策等因素综合影响的非线性行为方式,这其中包含了极大的不确定性和模糊性,同时又包含了某种必然性。利用有限的销售数据,选择合理的预测方法对药品销量进行预测对于医药流通企业合理安排进销存、减少经营成本具有重要的指导意义。本文首先介绍了所研究企业存在的问题以及准确销售预测的必要性,并介绍了预测领域的研究现状,对主要的预测方法进行了归纳;其次,针对药品销售过程中存在的线性、非线性、确定性、不确定性等问题,选取了较为合适的几个单项预测模型,介绍了相关的理论,并总结出各项模型的优缺点;再次,在各个单项模型的基础上提出了基于神经网络进行非线性组合的变权重组合预测模型—CAMB组合预测模型,给出其算法流程,并从理论上阐述了组合预测模型相对于单项预测模型的优性;最后,利用已有的药品销售数据,分别应用单项模型和CAMB模型进行预测,通过对模型预测结果的对比分析,验证了本文所提出的新模型的有效性和实用性。没有任何一种方法能保证在任何情况下都能得到较好的预测效果。本文所提出的CAMB组合预测模型将时间序列模型、云模型和神经网络模型结合起来,充分利用了各个模型的优点,将不同方面的有用信息综合起来,得到较优的组合预测模型,降低了单项模型出错的风险,为药品的销售预测提供了一种新的方法,提高了预测的准确度。
[Abstract]:The law of drug sales is not a simple linear behavior, but also a non-linear behavior influenced by market, economy, policy and other factors, which contains great uncertainty and fuzziness. At the same time, it contains some inevitability. Making use of the limited sales data and choosing reasonable forecasting methods to predict the drug sales volume has important guiding significance for pharmaceutical circulation enterprises to reasonably arrange the purchase, sale and storage, and to reduce the operating cost. This paper first introduces the existing problems of the enterprises studied and the necessity of accurate sales forecasting, and introduces the research status in the field of prediction, and summarizes the main forecasting methods; secondly, aiming at the linearity in the process of drug sales, Nonlinear, deterministic, uncertainty and other issues, selected a few more suitable single prediction models, introduced the relevant theories, and summed up the advantages and disadvantages of each model. On the basis of each single model, the variable weight combined forecasting model-CAMB combination prediction model based on neural network is proposed, and its algorithm flow is given. Finally, using the existing drug sales data, the single model and CAMB model are used to predict, and the results of model prediction are compared and analyzed. The validity and practicability of the proposed new model are verified. There is no way to guarantee a better prediction in any case. The CAMB combined forecasting model proposed in this paper combines the time series model, cloud model and neural network model, makes full use of the advantages of each model, synthesizes the useful information in different aspects, and obtains a better combination prediction model. It reduces the risk of single model error, provides a new method for drug sales prediction and improves the accuracy of prediction.
【学位授予单位】：华南理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：F721

【参考文献】